The present invention relates to a receiver for a mobile telecommunications system.

In WCDMA based mobile telecommunications such as the Third Generation (3G) system, a receiver such as a mobile telephone handset must perform a cell search to identify local cells when it is first switched on and periodically thereafter.

As cells local to the receiver are not necessarily synchronised, and the mobile terminal is not necessarily equidistant from each cell, one aspect of the cell search procedure involves synchronising the receiver to a cell. To facilitate this, primary and secondary synchronisation codes are transmitted by the cells as part of a transmitted signal destined for a receiver.

Figure 1 is a schematic illustration showing a frame of a signal transmitted by a cell of a 3 G system to a receiver. The frame, shown generally at 10, is made up of five subframes 12, with each sub frame 12 being made up of three slots 14. Each slot 14 comprises ten scheduling intervals (SIs) 16, with each scheduling interval being made up of 256 chips.

A primary synchronisation channel (P-SCH) code is transmitted in the first SI 18 of each slot 14, forming part of the transmitted signal. The P-SCH code is a 256 chip long sequence, with the same code value being repeatedly transmitted in the first SI 18 of a slot 14 of a transmitted signal. The sequence of code values is known to the receiver. Thus, by correlating the received signal with the known P-SCH sequence, the receiver can identify slot boundaries (i.e. when a slot starts) in the received signal, but the P-SCH code does not provide any information as to which of the fifteen slots 14 in a frame 10 starts at a given time.

To enable the receiver to identify the timings of individual slots 14 in the received signal, a secondary synchronisation channel (S-SCH) code sequence is used. The S-SCH code sequence is a sequence of fifteen S-SCH codes, with each code comprising 256 chips. One S-SCH code is transmitted in the first SI 18 of each slot 14, and thus it takes one complete frame 10 for the receiver to receive a complete S-SCH code sequence.

The 3GPP specification defines 64 unique S-SCH code sequences (uniquely associated with 64 'code groups'), each of which is fifteen codes in length, whilst there are sixteen possible different S-SCH codes. For example, code group 3 is as follows:

In other words, in slot 0 of the frame 10, S-SCH code number 1 is transmitted. S-SCH code number 1 is transmitted in slot 1, whilst code 2 is transmitted in slot 2, code 8 is transmitted in slot 3 and so on.

In order to identify which slot 14 of a frame 10 starts at a given point (in time) in the frame 10, the receiver correlates the received signal with each of the 16 possible S-SCH codes for each of the slots 14 in a frame 10. The most likely of the possible S-SCH codes may be selected for each slot 14 in the frame 10 on the basis of the results of the correlations performed, for example, such that a sequence of 15 S-SCH codes is produced. This sequence is then compared with each of the 64 unique S-SCH code sequences and 15 circular shifted versions of each of the 64 unique S-SCH code sequences, and the most likely of the pairs of S-SCH code sequences and time-shifts (e.g. code sequence 3, time shifted by 5 slots) is selected based upon the results of the comparison. The 64 unique S-SCH code sequences have a property that all of the sequences are different, and even if the sequence undergoes a cyclic shift, the shifted version of the sequence is not equal to any of the other sequences. Thus, any received sequence of 15 S-SCH codes can be used to determine which of the 64 unique code sequences is included in the signal received by the receiver, and which time shift must be used, thereby providing the receiver with information as to where the frame starts over a period of 15 slots. The approach described above allows a receiver to determine slot and frame boundaries in a received signal. However, this approach relies upon multiple correlation operations, which can be time consuming and computationally inefficient.

According to first aspect of the invention there is provided a receiver for receiving a signal containing a first code and a second code, the receiver comprising a first correlator for correlating the received signal with the first code, a second correlator for correlating the received signal with a plurality of candidate second codes and an adder for producing a plurality of values which are equal to the result of a correlation of the received signal with the sum of the first code and each of the plurality of candidate second codes.

According to a second aspect of the invention there is provided a receiver for receiving a signal containing a first code and a second code, the receiver comprising a first correlator for correlating the received signal with the first code, a second correlator and an adder for adding the result of the correlation of the first code to an intermediate result of a correlation of the received signal with a plurality of second codes such that the second correlator produces a plurality of values which are equal to the result of a correlation of the received signal with the sum of the first code and each of the plurality of second codes.

The plurality of values so calculated may be used in processing of the received signal, for example to determine which of the candidate second codes is most likely to be contained in the received signal.

In a receiver according to the present invention the effect of noise in the received signal is greatly reduced in calculating a value indicative of the correlation of the received signal with each of the candidate second codes. Thus, a more accurate evaluation of which of the candidate second codes is most likely to be contained in the signal is possible.

The second correlator may be configured to perform a hierarchical correlation of the received signal with inner and outer codes of the candidate second codes. By performing a hierarchical correlation in this way the number of correlation operations required can be reduced, thus reducing the computational burden on the receiver.

The first correlator may be configured to use intermediate results of a correlation of the received signal with an inner code of the candidate second codes in performing the correlation of the received signal with the first code. In this way unnecessary repetition of correlation operations can be avoided, thus further reducing the computational burden on the receiver.

The second correlator may be configured to perform a fast Hadamard transform to correlate the results of the correlations of the received signal with the inner code of the plurality of candidate second codes with the outer codes of the plurality of candidate second codes. This further reduces the computational burden on the receiver.

The adder may be configured to add the result of the correlation of the received signal with the first code to the results of the correlations of the received signal with the plurality of candidate second codes.

Alternatively, the adder may be configured to add the result of the correlation of the received signal with the first code to a result of a correlation of the received signal with an inner code of the plurality of candidate second codes.

The first correlator may be configured to perform a hierarchical correlation of the received signal with inner and outer codes of the first code.

The receiver may further comprise a processor for evaluating the plurality of values.

According to a third aspect of the present invention there is provided a method of processing a received signal, wherein the received signal contains a first code and a second code, the method comprising correlating the received signal with the first code, correlating the received signal with each of the plurality of candidate second codes and producing a plurality of values which are equal to the result of a correlation of the received signal with the sum of the first code and each of the plurality of candidate second codes.

According to a fourth aspect of the invention there is provided a method of processing a received signal, wherein the received signal contains a first code and a second code, the method comprising correlating the received signal with the first code, adding the result of the correlation of the first code to an intermediate result of a correlation of the received signal with a plurality of second codes and producing a plurality of values which are equal to the result of a correlation of the received signal with the sum of the first code and each of the plurality of second codes.

The correlation of the received signal may be performed as a hierarchical correlation of the received signal with inner and outer codes of the candidate second codes.

The correlation of the received signal with the first code may use intermediate results of a correlation of the received signal with an inner code of the candidate second codes.

A fast Hadamard transform may be performed to correlate the results of the correlation of the received signal with the inner code of the plurality of candidate second codes with the outer codes of the plurality of candidate second codes.

The result of the correlation of the received signal with the first code may be added to the results of the correlation of the received signal with the plurality of candidate second codes.

Alternatively, the result of the correlation of the received signal with the first code may be added to a result of a correlation of the received signal with an inner code of the plurality of candidate second codes.

A hierarchical correlation of the received signal with inner and outer codes of the first code may be performed. The method may further comprise evaluating the plurality of values.

Embodiments of the invention will now be described, strictly by way of example only, with reference to the accompanying drawings, of which:

Figure 1 is a schematic illustration showing a frame of a signal transmitted by a cell of a 3 G system to a receiver;

Figure 2 is a schematic illustration showing an architecture of a receiver for a telecommunications system;

Figure 3 is a schematic illustration showing one embodiment of an architecture which can be used to implement the invention; and

Figure 4 is a schematic illustration showing an alternative embodiment of an architecture which can be used to implement the invention.

Referring first to Figure 2, a receiver architecture is shown generally at 20. It will be appreciated that the functional blocks shown in Figure 2 are not necessarily representative of physical components of a receiver, but are used only for the purpose of illustrating the invention. Moreover, for reasons of clarity and brevity only those components of the receiver 20 which are relevant to the invention are illustrated, but it will be apparent to those skilled in the art that the receiver 20 comprises additional components.

The receiver 20 comprises an antenna 22 through which a signal can be received. Typically the received signal is an analogue signal which originates from a cell of a 3 G or other WCDMA based telecommunications system. The receiver 20 includes a sampler 24 and an analogue to digital converter (ADC) 26 for converting the received analogue signal into a sequence of digital samples. The received signal contains a primary synchronisation channel (P-SCH) code which is known to the receiver 20, with the known P-SCH code being stored in a memory 28 of the receiver 20. The received signal also contains a secondary synchronisation channel (S-SCH) code which is selected from a plurality (typically 16) of possible S-SCH codes. Each of the plurality of possible S-SCH codes is known to the receiver 20, which includes a memory 30 in which the plurality of possible S-SCH codes is stored.

Correlators 32 and 34 are provided for correlating the sampled received signal with the P-SCH code and each of the S-SCH codes. The results of these correlations are used in downstream processing of the received signal. For example, a selector 36 may be provided to select, on the basis of the results of the correlations, which of the S-SCH codes is the most likely to be present in the received signal.

In a known receiver, a sequence r of samples which represents the received signal is correlated with the known P-SCH code p to identify slot boundaries within a frame of the received signal. The P-SCH code p is a sequence of 256 bits, and is formed as an inner sequence a comprising sixteen bits having a value of either +1 or -1 and an outer sequence g, also comprising sixteen bits having a value of either +1 or -1. The P-SCH code /? is formed by multiplying the whole of the inner sequence a by each element of the outer sequence g to create the 256 bit sequence.

Thus, if

a = [ +1, -1, -1, +1, +1, -1, -1, +1, +1, -1, -1, +1, +1, -1, -1, +1] and

g = [ +1, -1, -1, -1, +1, -1, -1, -1, +1, -1, -1, -1, +1, -1, -1, -1], then

p = a x g = [ +a, -a, -a, -a, +a, -a, -a, -a, +a, -a, -a, -a, +a, -a, -a, -a]. The correlation of the received signal with/? may be accomplished in a 'brute force' manner, by performing 256 correlation operations to correlate the samples r with each of the bits of the known P-SCH code/?, i.e.

255

CorrelationValue = ∑ Ση p _t

However, this approach is time-consuming and computationally intensive.

An alternative approach is to make use of the hierarchical nature of the P-SCH code to reduce the computational burden and time required to calculate the correlation value by performing a hierarchical correlation. In this method, 16 bit sub-sequences of the sequence r are correlated with the inner sequence a of the P-SCH code to obtain 16 intermediate correlation values, and these intermediate correlation values are subsequently correlated with the outer sequence g of the P-SCH code, i.e.

CorrelationValue =

In this way, the number of correlation operations that must be performed to obtain the correlation value for the received samples r with the P-SCH code is greatly reduced, thus reducing the computational burden and time required to obtain the correlation value.

In a known receiver, the sequence r of received samples must also be correlated with known secondary synchronisation channel (S-SCH) codes to determine the position of individual slots within a received frame. A different S-SCH code is transmitted in each slot of the frame, and thus over the course of a frame, a sequence of fifteen S-SCH codes is transmitted. The 3GPP standard specifies sixteen possible S-SCH codes and 64 unique sequences of S-SCH codes.

Thus, when a signal is received, a known receiver correlates the sequence r of received samples with each of the possible S-SCH codes, and the results of this correlation may be used in downstream processing of the received signal, for example to identify which S-SCH code is contained in each slot of a received frame. Thus, every slot the receiver must perform sixteen correlation operations to determine which of the sixteen possible S-SCH codes is most likely to be contained in that slot of the received signal. Once an entire frame has been received, a sequence of most likely S-SCH codes is known and this can be compared to the 64 unique sequences and their 15 possible time-shifted versions by further correlation operations to determine which sequence was used in the received signal, and from this knowledge slot timing can be determined. The receiver can perform these operations in a 'brute force' manner, although this is computationally intensive, as it requires sixteen correlation operations to establish which of the sixteen S-SCH codes is most likely to be present in each of the fifteen slots of a frame, and a further 960 (=64 xl5) correlation operations once an entire frame has been received and the S-SCH codes identified to establish which of the 64 possible S-SCH code sequences, and which time-shift, is used in the received frame. Thus, 16 x 15 x 64 = 15360 correlation operations per frame are required if a 'brute force' approach is used.

However, the S-SCH codes, like the P-SCH code, have a hierarchical structure, and properties of this structure can be exploited by a receiver to reduce the computational complexity involved in processing the received signal, for example to identify which of the sixteen possible S-SCH codes is most likely to have been received.

Each of the sixteen S-SCH codes comprises an inner layer b of sixteen bits and an outer layer h, also of sixteen bits. Thus, each S-SCH code comprises 256 bits.

The inner layer b is common to all of the S-SCH codes and is related to the inner layer a of the P-SCH code by the following relationship

b = [α(0:7) , -α(8-15)]

In other words, the first eight bits of the inner layer b of each S-SCH are equal to the first eight bits of the inner layer a of the P-SCH, whilst the second eight bits of the inner layer b of the S- SCH are equal to the negative of the second eight bits of the inner layer a of the P-SCH. The outer layer h is different for each of the sixteen possible S-SCH codes, and comprises a sixteen bit sequence of positive and negative valued bits. Each S-SCH code is obtained by multiplying the inner layer b by each element of the respective outer layer h.

For example, suppose an outer layer ho comprises the following bit sequence

ho = [-1, +1, -1, +1, +1, -1, +1, -1, -1, +1, -1, +1, +1, -1, +1, -1]

Using the example of a given above,

b = [+1, -1, -1, +1, +1, -1, -1, +1, -1, +1, +1, -1, -1, +1, +1, -I]

Then the S-SCH code for this particular outer layer is

S ₀ = [-b, +b, -b, +b, +b, -b, +b, -b, -b, +b, -b, +b, +b, -b, +b, -b]

Known receivers employ hierarchical correlation operations and Hadamard transforms to reduce the computational complexity involved in processing the received signal, for example to identify which of the sixteen S-SCH codes is most likely to have been received in a slot of a received signal.

Thus, instead of correlating a 256 bit sequence of samples r which is assumed to be an S-SCH code which was received in a slot of a received signal with each of the sixteen possible S-SCH codes, a hierarchical correlation is performed by first correlating sixteen-bit sub-sequences of the sequence r with the inner sequence b of the S-SCH code to obtain sixteen intermediate correlation values, and these intermediate correlation values are subsequently correlated with the outer layers ho - his of each of the sixteen possible S-SCH codes, i.e.

CorrelationValue _n = K 'J This gives a correlation value for each of the sixteen possible S-SCH codes. By performing this hierarchical correlation, the number of individual correlation operations that must be performed is greatly reduced in comparison to the 'brute force' method.

To ease the computational burden further, a fast Hadamard transform may be employed to perform the sixteen correlation operations in parallel, as will be familiar to those skilled in the art.

The present invention improves on these known techniques by exploiting properties of the P- SCH code and the S-SCH codes so as to produce a correlation value in which the signal to noise ratio is improved, and this improved correlation value can be used in processing the received signal, for example to identify a most likely S-SCH code contained in the received signal.

The P-SCH and the S-SCH are known to be transmitted at the same time and with the same power. Moreover, the P-SCH code and the S-SCH codes cancel each other out half of the time because of the inter-relation of the inner code of the P-SCH code and the inner code of the S- SCH codes. This property is used in the present invention to obtain a better correlation value on which to base processing of the received signal, for example the determination of the S- SCH code used in the received signal, as will be explained below.

In known receivers, the sequence r is effectively correlated with each of the S-SCH codes in turn, i.e.

<R\ Yk> = <(X + Y + N)\ Yk>, , where R is a vector representing the sequence of samples r of the received signal, X is a vector representing the P-SCH code, Y _k is a vector representing the &th S-SCH code and N is a vector representing noise in the received signal.

As the P-SCH code and the S-SCH codes are orthogonal, they are un-correlated, i.e. <X\ Yu> = 0. Thus, <R\Y _k > = <Y\Y _k > + <N\Y _k > (1)

In the receiver of the present invention, the sequence r is effectively correlated with the sum of the P-SCH code and each of the S-SCH codes in turn, i.e.

CorrelationValue _k =< R | (X+Y _k ) > , where R is a vector representing the sequence of samples r of the received signal, X is a vector representing the P-SCH code and Y _k is a vector representing the &th S-SCH code.

The received signal contains the P-SCH code, one of the k S-SCH codes and noise. Thus,

R = X + Y + N , where Y is the unknown S-SCH code contained in the received signal and N is noise.

Correlating R with (X+ Y _k )

< R I (X+Y _k ) >=< (X + Y + N) I (X + Y _k )

= X ² +<Y X>+<N X> + <X\Y,, > + <Y Y ^l k,, > ^ + ' < ^N ^{J ϊ} \ IY ^± k,, >

As the P-SCH code X is orthogonal with each of the k S-SCH codes Y _k , <Y\X> = 0 and <^7*> = 0. Thus,

<R\(X + Y _k )>= X +<Y\Y _k > + <N\(X + Y _k )> (2)

This correlation value includes a constant term |X| and a noise term <N\(X+ Y _k )>, as well as the desired information regarding the correlation of the S-SCH code contained within the received signal with the sixteen known S-SCH codes (the term <Y\Y _k >). The signal to noise ratio in this correlation value is improved in comparison to known methods, as will be now be explained.

As is mentioned above, the P-SCH code X and each of the S-SCH codes Yk are orthogonal. The P-SCH code X and the S-SCH code 7 contained in the received signal R are transmitted at the same time cancel each other out for half of the time. The correlation of the noise with the sum of the P-SCH code Xand each of the S-SCH codes Y _k (for the 256 bit length of the P-SCH and S-SCH codes) can be written as

< N \ (X + Y _k ) >= Y _j n ₁ (x ₁ + y _h y ι=0

As the P-SCH and the S-SCH cancel each other out for half of the time, X ₁ + y _kl = 0 for half of the time. Thus, effectively

127

< N \ (X + Y _k ) >= ∑n _ι (x _ι + y _h ) * (3) i=0

Compare this to known methods in which the received signal is correlated with each S-SCH alone:

255

< N \ Y _k >= ∑ Σn _Jkι (4) i=0

The summation in equation (4) is carried out over 256 values, whereas the summation of equation (3) is only carried out over 128 values. However in equation (3), the amplitude of X ₁ + yki is twice the amplitude of y^ in equation (4), having a value of +/- 2 as opposed to +/-1. Therefore the average magnitude of the summation in equation (4) is the same as that of the summation in equation (3). In equation (2) the desired information regarding the correlation of the S-SCH code contained in the received signal with the sixteen known candidate S-SCH codes (the term <Y\ Yk>) is offset by a constant term \X\ ² . When Yk is equal to the S-SCH code Y contained in the received signal, <Y\ Y _k > = \Y\ ² and <R\(X + Y _k )> = \X\ ² + \ Y\ ² + <N\(X + Y _k )>, which is equal to 2\Y\ ² + <N\(X+Y _k )>, because \X\ ² = \Y\ ² as the P-SCH and the S-SCH are transmitted with the same power. In other words, the contribution of the desired signal in the correlation value of equation (2) is multiplied by two. Thus the effect of the noise in the received signal on the correlation operation which is performed during processing of the received signal, for example to determine which of the known S-SCH codes is contained in the received signal, is reduced in the receiver of the present invention.

The effect of correlating the received signal with the sum of the P-SCH code and each of the S-SCH codes in turn can be achieved in a number of ways. Figure 3 is a schematic illustration showing one possible receiver architecture.

In this embodiment, samples of the received signal produced by the ADC 26 are processed in two correlators 42 and 44. In correlator 42, the first eight received samples are correlated with the first eight bits of the inner code b of the S-SCH codes, whilst in correlator 44 the next eight received samples are correlated with bits 8 to 15 of the inner code b of the S-SCH codes. The inner code b is common to all sixteen of the S-SCH codes.

An adder 46 is provided to add the outputs of correlators 42 and 44 to provide the correlation of the received signal with the entire inner code b of the S-SCH codes as the output of the adder 46.

A further adder 48 subtracts the output of correlator 44 from the output of correlator 42. The inner sequence b of the S-SCH codes is equal to the first eight bits of the inner sequence a of the P-SCH code and the negative of the second sequence of eight bits of the inner sequence a of the P-SCH code (i.e. b = [α(0:7), -α(8:15)]), whilst the inner sequence a of the P-SCH code is equal to the first eight bits of the inner sequence b of the S-SCH codes and the negative of the second sequence of eight bits of the inner code b of the S-SCH codes (i.e. a = [b(0:7), - 6(8: 15)]). Thus, the output of adder 48 is the correlation of the received samples with the inner sequence a of the P-SCH code, which is efficiently calculated based on intermediate results for the correlation of the received signal with the inner code b of the S-SCH codes.

A sixteen- input calculation unit 50 is provided to implement a fast Hadamard transform on the output of the adder 46, to perform the correlation of the output of the adder 46 with the outer codes of each of the sixteen possible S-SCH codes, i.e. the output of the calculation unit 50 is <R\ Yk>.

A further calculation unit 52 is provided to perform the correlation of the output of the adder 48 with the outer code g of the P-SCH code, so that the output of the calculation unit 52 is <R\X>.

An adder 54 is provided to add the outputs of the calculation units 50, 52, so as to produce a correlation value for the sum of the P-SCH code with each of the sixteen possible S-SCH codes, which values can be evaluated by an evaluation unit 56 as part of the processing of the received signal, for example to determine which of the sixteen possible S-SCH codes is most likely to be present in the received signal. The output of the adder 54 is

<R\X> + <R\ Y _k > = <R\(X+Y _k )>.

Figure 4 is a schematic illustration of an alternative receiver architecture for achieving the effect of correlating the received signal with the sum of the P-SCH code and each of the S- SCH codes. In this example, samples of the received signal are correlated, in correlators 62, 64, with the first and second eight-bit sequences of the inner code b of the S-SCH codes. Adders 66 and 68 produce at their respective outputs a value for the correlation of the received signal with the inner code b of the S-SCH codes and a value for the correlation of the received signal with the inner code a of the P-SCH code, as in the example architecture shown in Figure 3. A calculation unit 70 is provided to correlate the value produced by adder 68 with the outer code g of the P-SCH code, thus producing as its output a correlation value for the correlation of the received signal with the P-SCH code, i.e. the output of the calculation unit 70 is <R\X>.

The output of the calculation unit 70 is added, in adder 72, to the first of sixteen outputs of adder 66, and the output of the adder 72 is provided to a first input of a sixteen-input calculation unit 74. The other fifteen inputs to the calculation unit 74 are provided by the fifteen unchanged outputs of the adder 66. The calculation unit 74 performs a fast Hadamard transform to correlate the input signal with the sixteen possible S-SCH outer codes. Thus, the calculation unit 74 produces a correlation value for the sum of the P-SCH code with each of the sixteen possible S-SCH codes, which values can be evaluated by an evaluation unit 76 as part of the processing of the received signal, for example to determine which of the sixteen possible S-SCH codes is most likely to be present in the received signal. The output of the calculation unit 74 is

<R\X> + <R\ Y _k > = <R\(X+Y _k )>.

The embodiments described above employ a hard decision based approach to determining which of the sixteen possible S-SCH codes is present in the received signal. However, it will be appreciated that other approaches are also possible. For example, a soft decision based approach may be used, in which the results of the correlations provided by the calculation unit 54, 74 over the course of a frame (15 slots) are stored in a memory as a sequence of fifteen "likelihood messages", each of which comprises a list of sixteen correlation results. The sequence of likelihood messages is then correlated with the each of the 64 unique S-SCH code sequences and the 15 circular shifted versions of the 64 S-SCH code sequences, to produce 960 (= 15 x 64) correlation results, each corresponding to one of the 64 S-SCH code sequences and their time-shifted versions. By evaluating the maximum of these correlation values the most likely pair of S-SCH sequence and time shift can be determined.

It will be appreciated that the functional blocks shown in Figure 3 and Figure 4 are representative only of functions carried out, and do not necessarily represent actual physical components of a receiver. The functional blocks may be implemented as separate components, such as processors, FPGAs or the like configured to perform the relevant operations, or may be configured as parts of a single processor, FPGA or the like. Alternatively, the functional blocks may be implemented in software executed by an appropriately configured processor, FPGA or the like.